puree 0.9.0 → 0.10.0

checksums.yaml

@@ -1,7 +1,7 @@
 ---
 SHA1:
-  metadata.gz: e16d74fee89f71aae7593434a3bb8041e6ae9451
-  data.tar.gz: 2180947af2af6b771305ac16712f447721f23f63
+  metadata.gz: 098c3fc6fe0dc9773261d20e4f2f8e021e9f8255
+  data.tar.gz: 49b91baf4a68f02a337b986c15f6fe35c67b4396
 SHA512:
-  metadata.gz: f56565ef90e7d266adf52c7f6a346e7e7d885e25f42064353f8b2840d6c31654d044e98d0a48f8cb677fb94f35ad47cfb26488b7db0b562cdbfe530bc93a3a6f
-  data.tar.gz: 360d2d041c5bb2c71739a245c4cfea4059c41c274cee2cae1289e3775bb9ab98c55297e59702e32137300e47da696d6cc17b5dd89fc4dcec261e766de46b2e47
+  metadata.gz: d3670c89f2adb4ff43f1c262ab10dbb140b4ab5438103a267f63524894292a760e41fc4af0839243cb61688608375b1c996a0960fe31e902b483861dddd1c723
+  data.tar.gz: c962958c92997235fc5730e310e15593cdd9d3ad6c7ab9db7c60db7ef61577c0abac86cd3ee359385354753a4ff899a2dd46d397413e0e624fc56f53bfc8d6b1

data/CHANGELOG.md

@@ -2,6 +2,12 @@
 All notable changes to this project will be documented in this file.
 This project adheres to [Semantic Versioning](http://semver.org/).
 
+## 0.10.0 - 2016-05-17
+### Added
+- Dataset metadata (associated, link, project, production as range, person for those without uuid, publication for all research outputs, publisher).
+### Fixed
+- Dataset metadata (description splitting, geographical stripping).
+
 ## 0.9.0 - 2016-05-16
 ### Added
 - Dataset metadata (production).

data/README.md

@@ -41,6 +41,7 @@ d.get uuid:     'xxxxxxxx-xxxx-4xxx-yxxx-xxxxxxxxxxxx',
 
 # Filter metadata into simple data structures
 d.access
+d.associated
 d.available
 d.description
 d.doi
@@ -48,6 +49,7 @@ d.file
 d.geographical
 d.keyword
 d.person
+d.project
 d.production
 d.publication
 d.temporal
@@ -123,8 +125,20 @@ An array of files.
 ]
 ```
 
+### link
+An array of links.
+
+```ruby
+[
+  {
+    "url": "http://www.example.com/~abc1234/xyz/",
+    "description": "An interesting description"
+  },
+]
+```
+
 ### person
-Contains an array of internal persons and an array of external persons.
+Contains an array of internal persons, an array of external persons and an array of other persons.
 
 ```ruby
 {
@@ -139,20 +153,65 @@ Contains an array of internal persons and an array of external persons.
     },
   ],
   "external"=>[
+  ],
+  "other"=>[
+      "name"=>{
+        "first"=>"Hal",
+        "last"=>"Roach"
+      },
+      "role"=>"Contributor",
+      "uuid"=>""
   ]
 }
 ```
 
+### project
+An array of projects associated with the dataset.
+
+```ruby
+[
+  {
+    "title": "An interesting project title",
+    "uuid": "xxxxxxxx-xxxx-4xxx-yxxx-xxxxxxxxxxxx"
+  },
+]
+```
+
 ### publication
-An array of related publications.
+An array of research outputs associated with the dataset.
 
 ```ruby
 [
   {
-    "type"=>"Journal article",
-    "title"=>"An interesting title",
-    "uuid"=>"xxxxxxxx-xxxx-4xxx-yxxx-xxxxxxxxxxxx"
+    "type": "Journal article",
+    "title": "An interesting journal article title",
+    "uuid": "xxxxxxxx-xxxx-4xxx-yxxx-xxxxxxxxxxxx"
+  },
+  {
+    "type": "Conference paper",
+    "title": "An interesting conference paper title",
+    "uuid": "xxxxxxxx-xxxx-4xxx-yxxx-xxxxxxxxxxxx"
+  },
+  {
+    "type": "Working paper",
+    "title": "An interesting working paper title",
+    "uuid": "xxxxxxxx-xxxx-4xxx-yxxx-xxxxxxxxxxxx"
+  },
+  {
+    "type": "Paper",
+    "title": "An interesting paper title",
+    "uuid": "xxxxxxxx-xxxx-4xxx-yxxx-xxxxxxxxxxxx"
   },
+  {
+    "type": "Dataset",
+    "title": "An interesting dataset title",
+    "uuid": "xxxxxxxx-xxxx-4xxx-yxxx-xxxxxxxxxxxx"
+  },
+  {
+    "type": "Chapter",
+    "title": "An interesting chapter title",
+    "uuid": "xxxxxxxx-xxxx-4xxx-yxxx-xxxxxxxxxxxx"
+  }
 ]
 ```
 

data/lib/puree/dataset.rb

@@ -8,17 +8,81 @@ module Puree
       super(:dataset)
     end
 
+    # Link
+    #
+    # @return [Array<Hash>]
+    def link
+      path = '//links/link'
+      xpath_result =  xpath_query path
+      data = []
+      xpath_result.each { |i|
+        o = {}
+        o['url'] = i.xpath('url').text
+        o['description'] = i.xpath('description').text
+        data << o
+      }
+      return data.uniq
+    end
+
+    # Publisher
+    #
+    # @return [String]
+    def publisher
+      path = '//publisher/name'
+      xpath_result =  xpath_query path
+      xpath_result ? xpath_result.text : ''
+    end
+
+    # Combines project and publication
+    #
+    # @return [Hash]
+    def associated
+      path = '//associatedContent//relatedContent'
+      xpath_result =  xpath_query path
+      data_arr = []
+      xpath_result.each { |i|
+        data = {}
+        data['type'] = i.xpath('typeClassification').text
+        data['title'] = i.xpath('title').text
+        data['uuid'] = i.attr('uuid')
+        data_arr << data
+      }
+      data_arr.uniq
+    end
+
+
+    # Project
+    #
+    # @return [Array<Hash>]
+    def project
+      associated_type('Research').uniq
+    end
+
+    # Publication
+    #
+    # @return [Array<Hash>]
+    def publication
+      data_arr = []
+      associated.each do |i|
+        if i['type'] != 'Research'
+          data_arr << i
+        end
+      end
+      data_arr.uniq
+    end
+
+
     # Title
     #
     # @return [Array<String>]
     def title
       data = node 'title'
+      data_arr = []
       if !data.nil? && !data.empty?
         data = data['localizedString']["__content__"]
-        data.is_a?(Array) ? data : data.split(',')
-      else
-        []
+        data.is_a?(Array) ? data_arr = data : data_arr << data
       end
+      data_arr.uniq
     end
 
     # Keyword
@@ -26,12 +90,12 @@ module Puree
     # @return [Array<String>]
     def keyword
       data = node 'keywordGroups'
+      data_arr = []
       if !data.nil? && !data.empty?
         data = data['keywordGroup']['keyword']['userDefinedKeyword']['freeKeyword']
-        data.is_a?(Array) ? data : data.split(',')
-      else
-        []
+        data.is_a?(Array) ? data_arr = data : data_arr << data
       end
+      data_arr.uniq
     end
 
     # Description
@@ -39,17 +103,17 @@ module Puree
     # @return [Array<String>]
     def description
       data = node 'descriptions'
+      data_arr = []
       if !data.nil? && !data.empty?
         data = data['classificationDefinedField']['value']['localizedString']['__content__'].tr("\n", '')
-        data.is_a?(Array) ? data : data.split(',')
-      else
-        []
+        data.is_a?(Array) ? data_arr = data : data_arr << data
       end
+      data_arr.uniq
     end
 
-    # Person, internal and external
+    # Person (internal, external, other)
     #
-    # @return [Hash<Array, Array>]
+    # @return [Hash]
     def person
       data = node('persons')
       persons = {}
@@ -60,6 +124,7 @@ module Puree
       end
       internal_persons = []
       external_persons = []
+      other_persons = []
       case data
         when Array
           data.each do |d|
@@ -72,6 +137,10 @@ module Puree
               person['uuid'] = d['externalPerson']['uuid']
               external_persons << person
             end
+            if !d.key?('person') && !d.key?('externalPerson')
+              person['uuid'] = ''
+              other_persons << person
+            end
           end
         when Hash
           person = generic_person data
@@ -83,36 +152,18 @@ module Puree
             person['uuid'] = data['externalPerson']['uuid']
             external_persons << person
           end
+          if !data.key?('person') && !data.key?('externalPerson')
+            person['uuid'] = ''
+            other_persons << person
+          end
       end
-      persons['internal'] = internal_persons
-      persons['external'] = external_persons
+      persons['internal'] = internal_persons.uniq
+      persons['external'] = external_persons.uniq
+      persons['other'] = other_persons.uniq
       persons
     end
 
-    # Publication
-    #
-    # @return [Array<Hash>]
-    def publication
-      data = node('relatedPublications')
-      publications = []
-      if !data.nil? && !data.empty?
-        # convert to array
-        data_arr = []
-        if data['relatedContent'].is_a?(Array)
-          data_arr = data['relatedContent']
-        else
-          data_arr[0] = data['relatedContent']
-        end
-        data_arr.each do |d|
-          o = {}
-          o['type'] = d['typeClassification']
-          o['title'] = d['title']
-          o['uuid'] = d['uuid']
-          publications << o
-        end
-      end
-      publications
-    end
+
 
     # Date made available
     #
@@ -129,7 +180,7 @@ module Puree
       data = node 'geographicalCoverage'
       if !data.nil? && !data.empty?
         data = data['localizedString']["__content__"]
-        data.is_a?(Array) ? data : data.split(',')
+        data.is_a?(Array) ? data.uniq : data.split(',').map(&:strip).uniq
       else
         []
       end
@@ -139,26 +190,15 @@ module Puree
     #
     # @return [Hash]
     def production
-      data = node('dateOfDataProduction')
-      Puree::Date.normalise(data)
+      temporal_range 'dateOfDataProduction', 'endDateOfDataProduction'
     end
 
+
     # Temporal coverage
     #
     # @return [Hash]
     def temporal
-      data = {}
-      data['start'] = {}
-      data['end'] = {}
-      start_date = temporal_coverage_start_date
-      if !start_date.nil? && !start_date.empty?
-        data['start'] = start_date
-      end
-      end_date = temporal_coverage_end_date
-      if !end_date.nil? && !end_date.empty?
-        data['end'] = end_date
-      end
-      data
+      temporal_range 'temporalCoverageStartDate', 'temporalCoverageEndDate'
     end
 
     # Open access permission
@@ -210,7 +250,7 @@ module Puree
 
         end
       end
-      docs
+      docs.uniq
     end
 
     # Digital Object Identifier
@@ -233,15 +273,19 @@ module Puree
     def metadata
       o = {}
       o['access'] = access
+      o['associated'] = associated
       o['available'] = available
       o['description'] = description
       o['doi'] = doi
       o['file'] = file
       o['geographical'] = geographical
       o['keyword'] = keyword
+      o['link'] = link
       o['person'] = person
+      o['project'] = project
       o['production'] = production
       o['publication'] = publication
+      o['publisher'] = publisher
       o['temporal'] = temporal
       o['title'] = title
       o
@@ -251,6 +295,8 @@ module Puree
 
     private
 
+
+
     # Assembles basic information about a person
     #
     # @param generic_data [Hash]
@@ -265,20 +311,57 @@ module Puree
       person
     end
 
+    # Temporal range
+    #
+    # @return [Hash]
+    def temporal_range(start_node, end_node)
+      data = {}
+      data['start'] = {}
+      data['end'] = {}
+      start_date = temporal_start_date start_node
+      if !start_date.nil? && !start_date.empty?
+        data['start'] = start_date
+      end
+      end_date = temporal_end_date end_node
+      if !end_date.nil? && !end_date.empty?
+        data['end'] = end_date
+      end
+      data
+    end
+
     # Temporal coverage start date
     #
     # @return [Hash]
-    def temporal_coverage_start_date
-      data = node('temporalCoverageStartDate')
+    def temporal_start_date(start_node)
+      data = node start_node
       !data.nil? && !data.empty? ? Puree::Date.normalise(data) : {}
     end
 
     # Temporal coverage end date
     #
     # @return [Hash]
-    def temporal_coverage_end_date
-      data = node('temporalCoverageEndDate')
+    def temporal_end_date(end_node)
+      data = node end_node
       !data.nil? && !data.empty? ? Puree::Date.normalise(data) : {}
     end
+
+    # Associated type
+    #
+    # @return [Hash]
+    def associated_type(type)
+      associated_arr = associated
+      data_arr = []
+      associated_arr.each do |i|
+        data = {}
+        if i['type'] === type
+          data['title'] = i['title']
+          data['uuid'] = i['uuid']
+          data_arr << data
+        end
+      end
+      data_arr
+    end
+
   end
+
 end

data/lib/puree/publication.rb

@@ -16,7 +16,7 @@ module Puree
       xpath_result =  xpath_query path
       data_arr = []
       xpath_result.each { |i| data_arr << i.text }
-      data_arr
+      data_arr.uniq
     end
 
     # Digital Object Identifier
@@ -44,7 +44,7 @@ module Puree
         doc['url'] = d.xpath('url').text
         docs << doc
       end
-      docs
+      docs.uniq
     end
 
     # Title
@@ -55,7 +55,7 @@ module Puree
       xpath_result =  xpath_query path
       data_arr = []
       xpath_result.each { |i| data_arr << i.text }
-      data_arr
+      data_arr.uniq
     end
 
     # Subtitle
@@ -66,7 +66,7 @@ module Puree
       xpath_result =  xpath_query path
       data_arr = []
       xpath_result.each { |i| data_arr << i.text }
-      data_arr
+      data_arr.uniq
     end
 
     # All metadata

data/lib/puree/version.rb

@@ -1,3 +1,3 @@
 module Puree
-  VERSION = "0.9.0"
+  VERSION = "0.10.0"
 end

metadata

@@ -1,14 +1,14 @@
 --- !ruby/object:Gem::Specification
 name: puree
 version: !ruby/object:Gem::Version
-  version: 0.9.0
+  version: 0.10.0
 platform: ruby
 authors:
 - Adrian Albin-Clark
 autorequire: 
 bindir: bin
 cert_chain: []
-date: 2016-05-16 00:00:00.000000000 Z
+date: 2016-05-17 00:00:00.000000000 Z
 dependencies:
 - !ruby/object:Gem::Dependency
   name: httparty